The invention is used for the detection of radiation in the range of wavelengths between 1 .mu.m and 1.7 .mu.m and more particularly at the wavelengths 1.3 .mu.m and 1.55 .mu.m, for example for applications in the field of telecommunication.
Such a photodiode is known in the prior art from the publication of Katsuya Hasegawa et al in "Extended Abstracts of the 16th International Conference on Solid State Devices and Materials", Kobe, 1984, p. 579-582, C-10-1, entitled "InGaAs PIN photodiode with Low-Dark Current".
This document describes a PIN photodiode constituted by a substrate of indium phosphide (InP) which is n.sup.+ doped on whose first surface a layer of indium phosphide (InP) is formed by epitaxy from the liquid phase, which layer is n.sup.- doped and whose concentration of charge carriers is of the order of 5.times.10.sup.15 cm.sup.-3. A layer of gallium indium arsenide which is n.sup.- doped and whose concentration of charge carriers is also of the order of 5.times.10.sup.15 cm.sup.-3 is deposited by the same process of epitaxy from the liquid phase on the layer of indium phosphide (InP) of the n.sup.- type which is etched so as to form a MESA structure. A layer of the p.sup.+ type is then formed by diffusion of Zn ions not only at the surface of the MESA structure, but also along the edges and along the entire circumference. The diffusion depth is 0.5 .mu.m at the surface of the MESA structure in the layer of InP. An ohmic annular contact is then formed by means of an alloy of Ti/Au at the surface of the MESA structure, while a metallic contact of Au/Sn/Au is formed on the surface of the substrate opposite to the MESA structure. Finally, a passivation layer of silicon nitride (Si.sub.3 N.sub.4) also serving as an antireflection coating is deposited at the center of the ohmic annular contact at the surface of the MESA structure and along the edges and the circumference of the latter.
These photodiodes have three advantages with respect to the prior art devices. First, since the outer edge of the p.sup.+ region is situated at the surface of the layer of InP, the leakage current is reduced. Second, the passivation by a layer of silicon nitride extended to the circumference of the diode permits the obtaining of higher quantum efficiency and a favourable lifetime of the diode. Third, the formation of a comparatively thick p.sup.+ layer around the MESA structure causes a guard ring of simplified construction to be constituted and causes the tunnel effect to be reduced.
The photodiodes obtained by means of semiconductors of the III-V group are actually of great industrial importance in optical or electro-optical devices operating in the range of wavelengths lying between 1 .mu.m and 1.7 .mu.m. For applications in the field of telecommunication, in which the wavelengths 1.3 .mu.m and 1.55 .mu.m are utilized, the requirements imposed, which are fairly stringent, tend to further improve the results obtained with the diodes described in the aforementioned document.
In fact, these known diodes have beside the qualities mentioned above a given number of disadvantages. More particularly, it is necessary to have a thickness of at least 0.5 .mu.m betweeen the surface on which an ohmic contact is formed--in this case the surface of the layer of InGaAs of the MESA--and the junction in order to avoid the degradation of the junction. However, since this distance is large, the absorption due to the InGaAs material of the p.sup.+ type is also large and consequently the efficiency is low. Moreover, the intrinsic capacitance of the diode is high. Finally, in the layers obtained by epitaxy from the liquid phase, it is difficult to control the n.sup.- doping and mostly an n.sup.- doping is obtained in the layer of InP which proves to be higher than the n.sup.- doping obtained in the layer of InGaAS..